MIT Joint Program on the
Science and Policy of Global Change
G
AL
B
O
L
SCIENCE
POLICY
CHANGE
M
I
T
Supplementarity: An Invitation to Monopsony?
A. Denny Ellerman and Ian Sue Wing
Report No. 59
April 2000
The MIT Joint Program on the Science and Policy of Global Change is an organization for research,
independentÊpolicy analysis, and public education in global environmental change. It seeks to provide leadership
inÊunderstanding scientific, economic, and ecological aspects of this difficult issue, and combining them into policy
assessments that serve the needs of ongoing national and international discussions. To this end, the Program brings
together an interdisciplinary group from two established research centers at MIT: the Center for Global Change
Science (CGCS) and the Center for Energy and Environmental Policy Research (CEEPR). These two centers
bridgeÊmany key areas ofÊthe needed intellectual work, and additional essential areas are covered by other MIT
departments, by collaboration with the Ecosystems Center of the Marine Biology Laboratory (MBL) at Woods Hole,
and by short- and long-term visitors to the Program. The Program involves sponsorship and active participation by
industry, government, and non-profit organizations.
To inform processes of policy development and implementation, climate change research needs to focus on
improving the prediction of those variables that are most relevant to economic, social, and environmental effects.
InÊturn, the greenhouse gas and atmospheric aerosol assumptions underlying climate analysis need to be related to
the economic, technological, and political forces that drive emissions, and to the results of international agreements
and mitigation. Further, assessments of possible societal and ecosystem impacts, and analysis of mitigation
strategies, need to be based on realistic evaluation of the uncertainties of climate science.
This report is one of a series intended to communicate research results and improve public understanding of
climate issues, thereby contributing to informed debate about the climate issue, the uncertainties, and the economic
and social implications of policy alternatives. Titles in the Report Series to date are listed on the inside back cover.
Henry D. Jacoby and Ronald G. Prinn,
Program Co-Directors
For more information, contact the Program office:
MIT Joint Program on the Science and Policy of Global Change
Postal Address: 77 Massachusetts Avenue
MIT E40-271
Cambridge, MA 02139-4307 (USA)
Location: One Amherst Street, Cambridge
Building E40, Room 271
Massachusetts Institute of Technology
Access: Telephone: (617) 253-7492
Fax: (617) 253-9845
E-mail: globalchange@mit.edu
Web site: http://web.mit.edu/globalchange/
Printed on recycled paper
Supplementarity: An Invitation to Monopsony?
A. Denny Ellerman and Ian Sue Wing1
Abstract
Article 17 of the Kyoto Protocol allows Annex B parties to meet their commitments by trading
greenhouse gas emissions reductions ÒsupplementalÓ to domestic emissions control. We demonstrate
that implementing supplementarity by imposing concrete ceilings on imports of allowances in a
market for tradable emissions rights gives rise to monopsonistic effects, even with price-taking
behavior by both buyers and sellers. We assess the importance of this finding for Annex B emissions
trading, in the context of the import and export provisions of the recent EU Proposal on
supplementarity. Our results show that the proposal would reduce efficiency, and could significantly
alter the distribution of the gains from trade in an Annex B tradable permits market.
Contents
1. Introduction ..........................................................................................................................................1
2. The Monopsonistic Effects of Supplementarity..................................................................................2
3. The EU Proposal on Supplementarity ...............................................................................................11
4. Conclusion ..........................................................................................................................................15
5. References ..........................................................................................................................................16
Appendix.................................................................................................................................................17
1. INTRODUCTION
Article 17 of the Kyoto Protocol allows Annex B parties to meet their commitments under the
Protocol by greenhouse gas (GHG) emissions trading so long as such trading is ÒsupplementalÓ
to domestic emissions control. Supplementarity in this context refers to the level of imports of
tradable emissions permits relative to abatement undertaken domestically.
The language of Article 17 clearly prohibits any party from relying entirely upon imported
permits to meet its commitment, but this limiting case is implausible. Each party faces a schedule
of identifiable GHG abatement opportunities, ranging in costs from near-zero to very high, and it
is reasonable to assume that parties would utilize these opportunities in a cost-effective manner.
Since every party would undertake some cheap domestic abatement, further elaboration of the
meaning of the supplemental provisions of the Kyoto Protocol could be deemed unnecessary.
1
Ellerman is executive director of the Joint Program on the Science and Policy of Global Change and senior lecturer
at the Sloan School of Management at the Massachusetts Institute of Technology. Sue Wing is a candidate for
the Ph.D. in Technology, Management and Policy at MIT and a research assistant with the Joint Program. We
are greatly endebted to Matti Vainio, Peter Zapfel, John Reilly, Erik Haites, and Henry Tulkens for valuable
comments on earlier versions of this paper. None are responsible for remaining errors. We also gratefully
acknowledge Henry TulkensÕ role in convening the session at the 1999 annual meeting of the European
Economic Association in Santiago de Compostela, Spain, which provided the initial impetus for the paper.
Funding for the research underlying this paper was provided by the corporate and government sponsors of MITÕs
Joint Program on the Science and Policy of Global Change, including particularly in this instance EPRI. As
always, much credit is due our colleagues in the Joint Program, who have been a continual source of stimulation.
Indeed, the Umbrella Group, consisting of Annex B parties that are not members of the European
Union (EU), have adopted this interpretation; however, the EU and its members maintain that
something more is required: an explicit quantitative limit, commonly referred to as a Òconcrete
ceiling.Ó To this end, the EU advanced a specific proposal to implement the supplemental
provisions of the Kyoto Protocol at the June 1999 meeting of the Subsidiary Bodies in Bonn.
This paper analyzes the effects of a limit on permit imports under the usual analytic
assumptions: namely, that all Annex B parties meet their Kyoto Protocol obligations and that
free trade in permits among these parties is a realistic possibility. The paper proceeds in two
parts. First, we address an aspect of supplementarity about which previous analyses have been
remarkably silent: the inherently monopsonistic effects of a restriction2 on permit imports. Then,
we present and explain the EU proposal, which is more complicated than the simple quantitative
limits that have been analyzed in earlier studies. In particular, we examine how various
provisions of the EU proposal affect the efficiency and equity attributes of emissions trading in
an Annex B market. A final section concludes, and a mathematical appendix presents proofs for
the essential points made in the body of the paper.
2. THE MONOPSONISTIC EFFECTS OF SUPPLEMENTARITY
A number of studies have examined the effect of restricting permit imports,3 and they present
broadly similar conclusions, namely,
¥ Misallocation of resources that increases the global cost of compliance with the Kyoto target,
¥ Reduction in the gains from emissions trading for exporters of permits, and
¥ Reduction in the cost of meeting importersÕ emissions-control obligations for some levels of
restriction.
The last point is an important one; however, several studies missed it entirely (Ellerman and
Decaux, and Ellerman, Jacoby and Decaux), and others observed the effect without drawing the
full implications. Both Bernstein et al. and Bollen et al. note the irony of results they obtain in
which the US, which opposes such restriction, is made better off by it, while the EU, which
advocates such restriction, benefits less or not at all. Criqui et al. produce results indicating that
all importers of tradable permits gain for some level of restriction. They describe a U-shaped
curve whereby the total cost for importers declines initially, as an import restriction of increasing
severity is imposed, and then rises until it reaches the higher cost level associated with autarky.
These results describe the effects of monopsony, yet surprisingly none of these studies
recognize that the implementation of supplementarity implies an exercise of monopsonistic
2
To avoid confusion in terminology, we use the term ÒrestrictÓ in all its forms when referring to the effect of the
supplementarity provisions, while ÒconstrainÓ and all its forms refers to the emission reduction obligations under
the Protocol. Thus, a restricted party will be a constrained party, but a constrained party is not necessarily restricted.
3
In particular, see the articles by Bernstein et al. (1999) and Bollen et al. (1999) in the special issue of the Energy
Journal devoted to the Kyoto Protocol. Criqui et al. (1999), Ellerman and Decaux (1998), and Ellerman, Jacoby,
and Decaux (1998) also address the effects of supplementarity in broader discussions of emissions trading.
2
power. Bernstein et al. dismiss their result as accidental (Òvery sensitive to assumptions about
baselines and a variety of elasticities and other parameters,Ó p. 247) and they stress that it should
not be taken Òas a strong guide to policy.Ó Criqui et al. correctly identify the effects as a more
general consequence of implementing restrictions, but cloud this central insight with a discussion
of compensation of exporters. Bollen et al. mention monopsony, only to dismiss the possibility
by arguing that Òrestrictions are not the result of monopsonistic optimizing behavior,Ó as if intent
were all that mattered. Other studies have discussed the potential exercise of market power in
permit markets, but the focus has always been on the relation of that potential to the initial
allocation of permits.4 None have pointed out that a concrete ceiling provides a coordinating
mechanism for restricting demand that could be as effective as overt collusion among buyers or
the exercise of market power by a significant importer.
Monopsonistic effects are illustrated most easily by assuming a competitive market in which
all agents are price takers. Such a market structure would be more likely to obtain if Annex B
partiesÕ assigned amounts were downloaded to agents within each economy and those agents
were free to trade within or across borders. With no restrictions on emissions trading, a free-trade
equilibrium would obtain in which the marginal costs of abatement would be equalized across
sources, the global cost of compliance would be minimized, and the gains from trade maximized.
Any effective restriction on permit imports will change total cost for the restricted party, as well
as for other market participants, and that change can be illustrated by the following differential.
dTCi = [ fi ( ai ) - p] dai + mi dp
(1)
For a party i, dTCi is the change in the total cost of emissions reduction, which may be
achieved by undertaking abatement ai (a small change in which is written dai) or importing
permits mi. The ith partyÕs marginal cost of domestic emissions control is determined by a
marginal abatement cost function fi(ai), and it purchases emissions allowances in the market at
price p (a small change in which is written dp).
The change in total cost of abatement depends on two effects. The first term on the right-hand
side describes the net change in domestic abatement cost and the second term describes the
change in the cost of imports. For any party that is restricted from its free trade equilibrium,
domestic abatement and the marginal cost of domestic control both increase. As a result, the first
term on the right hand side is positive. Appendix A1 shows that a binding restriction on any
partyÕs imports will reduce the total demand for permits and thereby the market price of permits,
so long as all other parties remain price-takers and face upward sloping marginal abatement cost
curves. The second term is therefore negative and the net impact of the import restriction on total
cost depends on the balance between these two terms. As illustrated graphically in Figure 1, the
restricted importer gains whenever the deadweight loss triangle ABC is less than the rectangle
CDEF, which represents the cost saving in remaining imports.5
4
5
For examples, see Westskog (1996); Bohm and Larsen (1994); and Baron (1999).
Note that the difference between domestic marginal cost and the market price for remaining permit imports will
create a rent that must be allocated in some manner.
3
Increasing the stringency of the
import restriction from nonbinding to autarky gives rise first
to a net saving, which increases
before it declines and eventually
disappears as rising domestic
abatement costs overwhelm the
savings on import costs, as
illustrated in Figure 2. This
sequence reflects the evolution of
the cost-saving and deadweight
loss components discussed more
Figure 1. Effect of Import Limit on Restricted Importer
fully in Appendix A3. With
convex marginal abatement curves the internal deadweight loss increases monotonically as the
restriction is tightened. As long as any permits are imported, the lower import price generates
cost savings that offset the internal deadweight loss. These savings increase initially and then
decline to zero as the autarkic limit is approached. The heavy line shows the net effect on the
total cost of compliance. Net savings are maximized at ar*, and any restriction up to the point
ar** will reduce importersÕ total cost of abatement relative to free trade.
Figure 2. Internal Deadweight Loss and Cost Savings due to Import Restriction
4
The differential in equation (1)
can also be used to explain the
consequences of import
restrictions for unrestricted
importers and exporters. Since
internal marginal abatement cost
will equal the market price of
permits for these parties, the first
right-hand-side term is zero.
Consequently, the effect of the
import restriction is determined
by the last term on the right-handside. For unrestricted importers
Figure 3a. Effect of Import Limit on Unrestricted Importer
(mi > 0) total cost is reduced,
while for exporters (mi < 0) the
negative cost (or benefit from
trade) is diminished. For discrete
changes in price and abatement,
the effect on unrestricted
importers is illustrated in
FigureÊ3a. The cheaper price for
permits leads these parties to
import more and they gain both
by the reduction in price for the
previous level of imports and by
the additional savings from
displaced domestic abatement.
Figure 3b. Effect of Import Limit on Competitive Exporter
AnÊunrestricted importer is the
demand-side equivalent of a non-cooperative fringe producer that produces more at the higher
price maintained by a dominant monopolist. In this context, unrestricted importers buy more and
bid the price up somewhat, thereby diminishing the price-depressing effects of the
supplementarity limit, as demonstrated in Appendix A2.
Figure 3b illustrates the case of the price-taking exporter. The net reduction in demand for
imported permits causes the price to fall and exporters to abate less (da). Exporters lose by the
amount, ABFCD, most of which (ABCD) is an income transfer to importers. The small triangle,
CEF, represents the global deadweight loss on the exportersÕ side of the market, and there will be
a corresponding deadweight loss on the importersÕ side of approximately equal magnitude.6
6
The exact amount of the loss on the importer side would depend upon the slope of importer marginal abatement
curves and the existence of unrestricted importers. If all importers were restricted and the slopes of importer and
5
ForÊimport restrictions that are not too stringent, it is easy to see that the magnitude of the
income transfer will be far larger than the global deadweight loss. Nevertheless, as the restriction
binds more tightly, the deadweight loss on both sides of the permit market increases until this
effect overwhelms the income transfer.
An interesting point is reached when the restricted import demand is less than the quantity of
hot air. The clearing price for permits will not be well defined at this point since permits have
great value to importers but the supply available at zero cost exceeds demand. In keeping with
the assumption of price-taking behavior, we assume a price of epsilon (equal to one dollar per
megaton (MT)) to cover transaction costs and to clear the market for the permits that can be
exported. AtÊthis point, the free market value of the hot air has been entirely transferred to the
importers and all the remaining deadweight loss occurs on the importer side. The term dp in
equation (1) becomes zero, and there are no further income transfers to mitigate the costs
incurred by importers as the restriction is tightened further.
The explanation of monopsonistic effects presented above can be made more concrete by
solving for the prices and quantities that clear the permit market. To do this we employ a set of
marginal abatement cost (MAC) curves, a prediction of business-as-usual (BAU) emissions, and a
set of Kyoto commitments for aggregate Annex B regions. The technique of solving for the market
equilibrium using MAC curves is described in Appendix A4. The MAC curves are generated for
the year 2010 from runs of the MIT Emissions Prediction and Policy Analysis (EPPA) model
version 3.0, as explained in Ellerman and Decaux (1998).7 EPPA is a recursive-dynamic
computable general equilibrium simulation of global economic activity, energy production and
use, and carbon emissions (Babiker et al., in preparation). It divides the world into 12 regional
economies, each represented by eight production sectors and inter-linked by trade in energy and
non-energy commodities. EPPA also predicts emissions of non-CO2 greenhouse gases, but these
are omitted from this analysis in order to maintain the transparency of the central results.8
exporter MACs were equal, then the total global deadweight loss would be twice the sum of the exporter
triangles corresponding to CEF. If some importers are unrestricted, the importer deadweight loss will be greater
since the lower price will induce unrestricted importers to take up some of the restricted demand. In fact, the
slopes of the MACs are not greatly different, as indicated by the values for the parameter b in the next footnote.
7
The MACs are constructed by imposing progressively stringent proportional constraints on carbon emissions in
Annex B regions, and by using a least-squares procedure to fit a constant elasticity function to the model results
in price-abatement space. The approximations take the form
mci = a i ai b i , whose coefficients for the six
AnnexÊB regions are shown in the table below: The six regions are the United States, Japan, the 15 states of the
European Union, other OECD nations, former Soviet Union, and European Economies in Transition.
a
b
8
USA
0.0113
1.5682
JPN
0.1164
1.6809
EEC
0.0373
1.4903
OOE
0.1444
1.5914
FSU
0.0643
1.3918
EET
0.1883
1.5511
The net effect of including the non-CO2 gases on both permit imports and the incidence of import restrictions is
unclear, since their inclusion not only augments cost-effective domestic abatement options but also increases the
quantity of GHGs that parties are required to control. As shown by Reilly et al. (1999), inclusion of these gases
leads to lower autarkic marginal costs of abatement and lower total costs of meeting the Kyoto commitments.
6
The prediction of BAU carbon emissions is taken from the reference scenario in EPPA 3.0
and the cost implications of the Kyoto commitments using the EPPA-generated MACs and BAU
emissions are given in Table 1. The EPPA reference scenario is a relatively high emissions
forecast, which results in greater abatement and higher costs than some others. The recently
published reference forecast in the 1999 International Energy Outlook (hereafter, IEO99)
published by the Energy Information Administration of the US Department of Energy
(USDOE,Ê1999) is also included in Table 1 for comparison.
Table 1. Cost Projections with EPPA MACs (EPPA and IEO99 Reference Emissions)
EPPA Reference
IEO99 Reference
Baseline Emissions (MT Carbon)
Annex B
USA
Japan
Western Europe
Other OECD
Eastern Europe
FSU
4,762
1,850
351
1,189
309
275
788
4,344
1,790
322
1,021
275
270
666
b
94
256
238
208
201
-
35
217
126
71
98
c
32.5
38.2
5.6
20.0
5.5
-7.4
-29.5
Kyoto Reduction (MT Carbon)
6.1
15.1
1.6
3.5
1.4
-2.3
-13.1
447
538
64
159
60
-50
-324
Autarkic Total Cost (Billion 1995 US$)
102.4
59.5
8.3
27.4
7.3
-
Free Trade Total Cost (Billion 1995 US$)
Annex B
USA
Japan
Western Europe
Other OECD
Eastern Europe
FSU
IEO99 Reference
a
865
598
93
327
94
-45
-202
Autarkic Marginal Cost (1995 US$)
Annex B
USA
Japan
Western Europe
Other OECD
Eastern Europe
FSU
EPPA Reference
55.2
45.4
3.0
4.5
2.3
-
Free Trade Import Share (%)
47
42
41
38
-
69
54
38
48
-
a
Notes: Negative entries indicate predicted emissions below assigned amount, or Òhot air.Ó
b
Annex B entries are the clearing prices for an Annex B market with trading.
c
Negative entries denote net benefits, or export revenues less abatement cost.
Several points that emerge from the comparison of these two forecasts should be kept in mind
in an analysis of supplementarity. First, relatively small changes in predicted emissions can have
large effects on partiesÕ abatement requirements and the characteristics of an emissions trading
market. Annex B emissions are only 9% less in the IEO99 forecast than in the EPPA 3.0
reference case (4.34 GT vs. 4.76 GT), but the amount of abatement required to meet the Kyoto
commitments is half as much (0.45 GT vs. 0.87 GT) and the costs are less by the same
proportion or more. Since the ceiling is fixed, the difference between the forecasts is fully
7
reflected in the amount of required abatement, which is the primary determinant of cost. Second,
since lower growth implies less required abatement, lower permit prices, and greater reliance on
imported permits, the probability that any given concrete ceiling would be binding increases with
lower growth forecasts. Third, assumptions about the growth in emissions have a much greater
effect on total costs when trading is assumed than otherwise. For instance, when no trading is
assumed, total Annex B costs are about half as much with the lower growth IEO99 forecast than
with the higher growth EPPA reference case. When unrestricted trading is assumed, total costs in
the low growth forecast are one-sixth the amount of the higher growth case. Trading makes hot
air available and there is more of it when exportersÕ emissions grow slowly.
Finally, it should be noted that both of these forecasts agree on one new feature that reflects
emissions growth since 1990. Earlier analyses typically showed that Japan would face
considerably higher marginal cost of abatement without trading than other Annex B regions, and
therefore that it would be the party gaining the most from emissions trading, as well as from
certain levels of restriction. The US now occupies that position, or shares it with Japan
depending on the forecast, due to the greater rate of economic and emissions growth during the
1990s in the US compared to Japan, the EU or most other OECD countries.9
Figure 4 uses the curves first developed by Criqui et al. (1999) to show the effect on total cost
of the Kyoto reduction as the import restriction moves from none at all to complete prohibition.
For ease of analysis and exposition, we define the ceiling as the minimum percent of domestic
Figure 4. Effect of Import Restriction on Total Cost (EPPA Reference)
9
When equal percentage reductions are required, Japan still has the highest marginal cost of abatement; however,
JapanÕs lower economic and emissions growth in the 1990s has substantially lessened the reduction required of
Japan as a percentage of BAU emissions, when compared for instance with the US (20% vs. 30% in the IEO99
forecast and 26% vs. 32% in the EPPA reference case).
8
abatement required. This percent ranges from 0% representing no restriction (free trade) to 100%
representing an absolute prohibition on international permit trading (autarky). Thus, a limitation
restricting permit imports to no more than 25% of the required abatement can be expressed
equivalently as a requirement to abate at least 75% domestically. In Figure 4, the total cost of the
Kyoto reduction associated with given levels of restriction is expressed relative to the total cost
under free trade, or in the case of the exporting regions, the total benefit with free trade.
For importers, the consequences of a concrete ceiling fall into one of three categories. The
first includes levels of restriction that would not bind on any importing region, such as, in the
EPPA 3.0 reference forecast, a ceiling requiring up to 53% domestic abatement (prohibiting
imports of more than 47% of the total requirement). Any concrete ceiling within this interval will
have no effect on costs since all importing parties would be abating at least this much
domestically. The second category contains all levels of restriction that result in lower total cost
for importers than that associated with free trade. This second subset will be least inclusive for
the importer facing the highest autarkic marginal cost and the most inclusive for the importer
with the lowest autarkic marginal cost. For instance, using the EPPA 3.0 reference forecast, this
range begins at 53% required domestic abatement for all parties, but ends at 84% for the US and
at 90% for the OOE. The third subset includes all levels of restriction that increase importersÕ
total cost above that associated with free trade. This subset is the complement to the second, and
it runs from 84% to 100% for the US and from 90% to 100% for the OOE.
The experience of the exporting regions is the same as that for importing parties over the first
intervalÑno effectÑbut it is quite different as soon as the first importing party is actually
restricted. Thereafter, the benefits associated with free trade are diminished continually until the
point is reached, in this case when 80% domestic abatement is required, at which the demand for
imports is less than the available hot air. At this point, exporters undertake no abatement, and,
with price-taking behavior, they are assumed to sell whatever hot air is demanded for a token
sum of one dollar per MT.
The reduction in the gains from trade from import restriction and the distribution of those
gains are illustrated on Figure 5, which plots the aggregate gains by importers (the four OECD
regions) and exporters (FSU and EET) along the horizontal and vertical axes, respectively. The
point labeled free trade identifies the maximum gains from trade and the distribution of those
gains resulting from the Kyoto allocation of assigned allowable amounts (AAUs) and the EPPA
3.0 reference forecast. The dashed line passing through the free trade point is the emissions
trading equivalent of a budget line representing all possible allocations of the maximum gains
from free trade between importers and exporters.10 Any point lying between this line and the
origin indicates a loss of efficiency, which can be measured by the perpendicular distance from
that line. The 45° line indicates all points at which the gains from trade are evenly split between
10
Given the EPPA 3.0 reference forecast and the assumption of competitive market structure, any other point on this
line could be attained by a redistribution of assigned amounts between importers and exporters in what is a simple
illustration of the Coasian Theorem applied to GHG emissions trading. For instance, if the allocation of AAUs
agreed to at Kyoto resulted in greater allocations to the OECD economies and commensurately fewer to the FSU
and Eastern Europe, the free trade point would have been moved to the right along this line, and vice versa.
9
Figure 5. Effect of Restriction on Importer and Exporter Gains from Trade (EPPA Reference)
importers and exporters, which is used here to depict equity. Equity is of course a notoriously
hard concept to quantify and we do not wish to suggest that an equal distribution of the gains
from trade among importers and exporters is anything more than a heuristic concept of equity.
Nevertheless, any more appropriate definition of equity would allocate the gains between
importers and exporters in a manner that could be plotted in this space. Departures from this line
would then provide a measure of the equity effects of the restriction.
The curve in Figure 5 indicates the effect of an increasingly more stringent, uniform import
restriction on the aggregate gains from trade and the distribution of those gains between
importers and exporters. The free trade point prevails for any uniform restriction mandating up to
53% domestic abatement. Restrictions of 55% and 60% domestic abatement result in little loss of
efficiency, but the gains from trade are increasingly redistributed in favor of importers. The
maximum gain for importersÑand their least cost of complying with the Kyoto ProtocolÑwill
be achieved with a limit requiring 70% domestic abatement. Even tighter restrictions on
importers move the locus of gains back towards the origin and eventually to the loss of all gains
from emissions trading, which is the definition of autarky. The vertical line XX separates the
curve into ranges for which importers gain (to the right) or lose (to the left) relative to free-trade.
In this case, aggregate importer gains are possible over a wide range of restriction, from 53% to
slightly over 85% required domestic abatement.
The specific results shown in Figures 4 and 5 reflect the particular forecast used and the
allocation of AAUs in the Kyoto Protocol. Different forecasts and allocations would lead to a
different free trade equilibrium point, which may be closer or farther from the origin and closer
to one axis or the other than depicted in Figure 5. Also, depending on the forecast and the
allocation, the point at which a concrete ceiling would bind would be higher or lower. What
should be clear, however, is that, as soon as the import restriction binds, the monopsonistic
effects shown in Figures 4 and 5 will appear. A restriction on imports can reduce the total costs
for importers over a considerable interval of restriction because of the monopsonistic
redistribution of the (reduced) gains of trade that is inherent in a restriction on demand.
10
3. THE EU PROPOSAL ON SUPPLEMENTARITY
At the June 1999 meeting of the Subsidiary Bodies, the EU advanced a specific proposal to
implement the supplemental provisions of the Kyoto Protocol.11 This proposal is considerably
more complicated than the single uniform limit used in the preceding section and in other
analyses of supplementarity. Therefore, we quote it in full.
ÒNet acquisitions by an Annex B Party for all three Kyoto mechanisms together must not
exceed the higher of the following alternatives:
¥ 5% of its baseline emissions multiplied by five and its assigned amount (over five years)
divided by two, or
¥ 50% of the difference between its annual actual emissions in any year of the period from
1994 to 2002, multiplied by five, and its assigned amount.
ÒNet transfers by an Annex B Party for all three Kyoto mechanisms together must not exceed:
¥ 5% of its baseline emissions multiplied by five and its assigned amount divided by two.
ÒHowever, the ceiling on net acquisitions and on net transfers can be increased to the extent
that an Annex B Party achieves emission reductions larger than the relevant ceiling in the
commitment period through domestic action undertaken after 1993, if demonstrated by the
Party in a verifiable manner and subject to the expert review process to be developed under
Article 8 of the Kyoto Protocol.Ó
There are three distinct components: a restriction on imports, a restriction on exports, and the
however clause. The last two have a significant effect on the monopsonistic consequences of
supplementarity.
The conventional import component of the EU proposal would limit a partyÕs imports of
tradable permits during the first commitment period to no more than the greater of the two
formula quantities. These formulae may appear complicated, but they are a reasonable attempt to
deal with an inherent problem in implementing an import restriction. Analysts typically model a
restriction as a percent of the reduction required by the Kyoto Protocol, but the counterfactual
emissions implicit in such a formulation are never observed. The EU proposal avoids this
problem by referring to quantities already agreed upon or to emissions that will be observed
before the proposed restriction becomes applicable. For any given forecast, such as the EPPA 3.0
reference case, the EU limits for both imports and exports can be readily calculated and
converted to percentages of the required emission reduction in 2010, as shown in Table 2. The
EU import limit would not be uniform among parties, as assumed in Figures 4 and 5, but would
vary depending on growth of emissions since 1990. Using the EPPA 3.0 reference forecast,
domestic abatement would be required for 72Ð77% of an importing partyÕs reduction
requirement, or equivalently, imports would be limited to 23% to 28% of the requirement.
11
Submission by Germany on Behalf of the European Community, its Member States, and Croatia, Bulgaria, Czech
Republic, Estonia, Hungary, Latvia, Lithuania, Poland, Slovak Republic and Slovenia on Emissions Trading
(Art. 17 KP); Principles, Modalities, Rules, and Guidelines for the Mechanisms Under Articles 6, 12, and 17 of
the Kyoto Protocol, Note by the Secretariat, Addendum (FCCC/SB/1999/MISC.3/Add.3).
11
Table 2. Calculation of EU Supplementarity Limits in EPPA 3.0 Reference Case
5% Formula Amount (MT)
a
50% Formula Amount (MT)
Greater of the two (MT)
Required emission reduction (MT)
% of required reduction
USA
EEC
JPN
OOE
EET
FSU
65
169
169
598
28%
43
75
75
327
23%
13
27
27
93
28%
11
24
24
94
26%
16
-
50
-
a
Note: Based on actual emissions through 1998 and interpolated amounts for later years.
The export component of the EU proposal would limit a partyÕs permit exports to the first of
the two formulae applying to imports, or to a total of 66 MT of carbon. Restricting permit
exports as a means of implementing supplementarity is a novel interpretation of the Kyoto
Protocol, but it does reflect a concern often expressed about the export of hot air. Moreover, it
might be argued that an export limit provides an indirect way of implementing supplementarity
since the higher market price associated with restricted exports of hot air reduces permit imports
from what they would be otherwise.12 The proposed export limit also effectively redefines the
assigned amounts for would-be exporters in the Kyoto Protocol, at least to the extent that they or
others anticipated trading hot air in the first commitment period in excess of the proposed limit.
The however clause raises the proposed ceilings on both imports and exports to the extent that
an Annex B party undertakes verifiable domestic abatement. An importing party can exceed the
formula quantities to the extent it can demonstrate domestic abatement greater than that amount.
Thus, the clause effectively raises the import ceiling, and reduces the domestic abatement
requirement, to 50% of the emission reduction requirement. By implication, a party importing
more than the formula amount but meeting less than 50% of its requirement by domestic
abatement would receive credit for imports equal only to the level of domestic abatement. For
instance, a party otherwise in compliance that had imported permits to cover emissions 60 MT
over the Kyoto limit but that could demonstrate only 40 MT of domestic abatement, would
presumably be judged 20 MT short of its commitment.
On the export side, the however clause operates in a similar manner. Annex B parties could
export more than the formula amount if verifiable domestic abatement by the exporting party can
beÊdemonstrated. Assuming that hot air can be distinguished from real abatement, unconstrained
parties would be limited to exporting only the formula amount of hot air, but they would not
otherwise be restricted. Constrained parties (those with no hot air), which might export due to low
marginal cost, would be protected up to the formula amount from challenges based on additionality,
but for exports in excess of the formula amount, verifiable abatement would need to be shown.
12
Although the export limit will result in greater Annex B abatement during the first commitment period, the
provision leads to no greater cumulative abatement given the ability of any party to bank unused assigned
amounts against presumed reduction requirements in later commitment periods by Article 3.13 (See Manne and
Richels, 1998, for an analysis of the relative benefit of banking versus selling in the first commitment period.)
12
The loosening effect of the however clause is entirely dependent on the possibility of proving
that abatement is real. If verifiable abatement could be demonstrated with relatively little cost,
the however clause would operate as intended: to relax the otherwise very restrictive formula
limits on emissions trading. Such is its effect in the first-best world of model simulations in
which all abatement is real and trading is frictionless; however, in a less perfect worldÑwhere
the counterfactual emissions required to establish verifiable reductions can never be definitively
provenÑthe added test invites mischief that would frustrate the intent of the clause. Importers
who would use the clause will find themselves embroiled with their challengers in an ultimately
irresolvable duel of competing model projections concerning what domestic emissions would
have been had no domestic actions been undertaken to comply with the Kyoto Protocol.
Exporters who would take advantage of the however clause will find themselves in the same
dilemma for the distinction between hot air and real abatement depends similarly on establishing
what exporter emissions would have been without any abatement in response to the Kyoto
Protocol. At best, the added test needlessly increases transaction costs; at worst, it will prove
theÊhowever clauseÕs promise of relief to be illusory.
The several components of the EU proposal have a large effect on the efficiency and equity
aspects of the distribution of the gains from emissions trading. Four points on Figure 6 can be
used to illustrate the effects of the import and export limits alone and of the however clause.
TheÊtwo points labeled Òw/o however clauseÓ indicate the division of the gains between
exporters and importers when either the import or export limit is applied. The import limit
diminishes the global gains from trade by about 20% relative to the unfettered market
equilibrium, but the income transfer is such that, from the standpoint of the importers, a nearly
optimal monopsonistic restriction is achieved. The export limit without the however clause
Figure 6. Importer and Exporter Gains From Trade with EU Limits (EPPA Reference)
13
would restrict the supply of cheap abatement from the non-constrained Annex B parties to
66ÊMT, but that small amount is sold at a very high price: high enough in this case to cause two
of the constrained Annex B regions to become small exporters. The two non-constrained regions,
the FSU and eastern Europe, are better off than with the import limit, but the importers are much
worse off, and nearly all the potential gains from emissions trading have been given up.
When the however clause is present and operating as intended, it moves both of these points
toward the free trade equilibrium. In the EPPA 3.0 reference case with the import limit, no
importing party would be restricted when the however clause operates since all constrained
parties would abate at least 50% of their emission reduction requirements domestically. As a
result, the point representing the import limit with the however clause is the same as that for the
free trade equilibrium. When the export limit applies, the however clause allows exporting
parties to sell as much real abatement as importers are willing to buy. All parties gain from the
relaxation of the export limit, as shown by the point labeled Òexport limit with however clause.Ó
In fact, this point represents the efficiency and equity implications of the full EU proposal
(bothÊlimits with however clause operating) when it is applied in an Annex B market. In this
instance, only the export limit is binding since the clearing price is high enough to cause all
importing regions to abate more than 50% of their emission reduction requirement domestically.
Compared to free trade, the proposal reduces total surplus by about 28% to $50 billion, roughly
60% of which would accrue to exporters.
The four outcomes illustrated by Figure 6 depend greatly in magnitude and exact location
upon the BAU emissions prediction, but the relative positions shown here will hold for all
reasonably plausible forecasts of Annex B emissions during the first commitment period. For
instance, in the IEO99 forecast, some parties import more than 50% of the reduction requirement
with free trade so that the import limit with the however clause would not coincide with the free
trade equilibrium using this forecast. In general, when the EU import limit binds, the locus of
gains is moved to the southeast into the region showing monopsonistic effects, and when the EU
export limit binds, the locus is moved mostly toward the origin. For both limits, the however
clause operates to place these points nearer, or even at, the free trade equilibrium.
The point representing the full EU proposal (export limit with however clause) can be
usefully compared to the three points within the oval on Figure 6, which represent solutions in
which the FSU is assumed no longer to be a price taker but to exercise its potential for monopoly
power. Each dot within the oval reflects a different assumption about importer behavior: no
restriction, the EU limit on imports without the however clause, and the EU limit on exports with
the however clause.13 The exercise of monopoly power by the largest supplier always entails a
loss of efficiency, but the principal effect is to redistribute the gains from buyers to sellers, as can
be seen by comparing each point to the corresponding position where the FSU is assumed to be a
13
The import limit with the however clause is that same as being unrestricted in this forecast, and the export limit
without the however clause removes all exporter discretion.
14
price taker.14 Although suppliers can always gain by exploiting market power, if they are unable
to do so, the proposed EU export limit approximates the result by providing the requisite
mechanism for restricting supply. By so doing, the EU export limit succeeds in transforming the
unwelcome invitation to monopsony into a self-abnegating invitation to monopoly, at least in an
Annex B market.15 But this monopolistic result depends critically on whether the however clause
works as intended. If it does not the export limit will still be the binding constraint, and whether
the FSU is a price taker or monopolist will be a moot point, for any exports beyond this limit will
be open to question based on exportersÕ counterfactual emissions. In this case, the outcome will
be the point closest to the origin where there is little trading, and little gain for anyone.
4. CONCLUSION
The title to this paper indicates the conclusion. The supplementarity condition in the Kyoto
Protocol is an invitation to the exercise of monopsonistic power, whether intended or not.
Monopsonistic effects require only that demand be restricted, not that the restriction be optimal
in any sense. Also, these effects occur with price-taking behavior by both buyers and sellers so
long as an effective coordinating or restraining mechanism is in place. A concrete ceiling on
permit imports provides such a mechanism, and with it, all buyers will be better off over some
range of restriction, although most assuredly there is a point beyond which further restriction
results in higher costs for importers than under free trade. Thus, the unexpected results noted by
many analysts are not accidental features of import restriction, but unavoidable consequences.
The EU proposal avoids the monopsonistic effects that are addressed in this paper, but it does
so only by calling into question the basic elements of the agreement among the parties at Kyoto.
Emissions trading and hot air were included in the Kyoto Protocol to lower the costs of the
emissions reductions sufficiently to gain agreement and encourage parties to take the first steps
toward averting the threat of dangerous climate change. Supplementarity was also included,
raising the questionÑat least implicitlyÑof what was agreed upon. This provision can be
interpreted as stating the obvious about the necessarily supplemental nature of imported permits
in any serious implementation of Kyoto. Unfortunately, and as if to illustrate the proposition that
seemingly innocuous provisions are rarely such, the EU proposal goes well beyond the simple,
relatively non-binding restriction that might have been imagined. The export limit effectively
vitiates whatever role the excess AAUs were intended to have. And setting aside hot air, the
beneficial effects of emissions trading for both importers and exporters depend entirely on the
workability of the however clause, which hinges on yet another seemingly innocuous provision,
that verifiable abatement be shown. One might understandably question whether there was any
agreement on emissions trading and hot air at Kyoto.
14
Conversely, the dots within the oval indicate the effect of the proposed EU limits if monopolistic behavior is
assumed from the beginning. Neither restriction improves importer gains.
15
The import limit would not be redundant in a global market because of the abundant low-cost supplies from
non-Annex B exporters. In this market, the export limit mitigates the impact of monopsonistic effects on
non-Annex B suppliers.
15
The EU proposal on supplementarity is a well-intentioned and in many ways clever
interpretation of this worryingly vague provision of the Kyoto Protocol, but it is also illustrates
how good intentions and bad assumptions can make things worse. The good intention is the very
goal of the Kyoto Protocol: that parties undertake real abatement. The bad assumption is that
emissions trading could lead a party to avoid undertaking any real abatement at all. It is hard to
imagine how this latter circumstance could occur, other than by permits that are fraudulent, in
which case the problem is fundamental and more severe than can be addressed by restrictions on
emissions trading. In the end, the appropriate distribution of emissions reductions achieved
domestically versus abroad and now versus later (e.g., hot air) is a decidedly secondary concern
to the more important issue of creating the monitoring and accounting mechanisms to ensure that
all traded permits are excess to the issuing countryÕs needs. In focusing on such concerns, and in
introducing equally worrisome distinctions, the EU proposal threatens to become a distracting
side-show from the main event, which is to build a workable and effective international structure
to encourage parties to undertake actions to reduce emissions of greenhouse gases.
5. REFERENCES
Babiker, Mustafa H., J.M. Reilly, R.S. Eckaus, I. Sue Wing, 2000, EPPA 3.0 Documentation,
mimeograph (in preparation).
Baron, Richard, 1999, Market Power and Market Access in International GHG Emissions Trading.
IEAÊInformation Paper, International Energy Agency (Paris), October.
Bernstein, Paul, W.D. Montgomery, T.F. Rutherford and G. Yang, 1999, Effects of Restrictions on
International Permit Trading: The MS-MRT Model, The Energy Journal, Special Issue: The Costs
ofÊthe Kyoto Protocol: A Multi-model Evaluation, pp. 221-256.
Bohm, Peter and B. Larsen, 1994, Fairness in a Tradeable-Permit Treaty for Carbon Emissions
Reductions in Europe and the Former Soviet Union, Environmental and Resource Economics,
4(3):Ê219-39.
Bollen, Johannes, A. Gielen and H. Timmer, 1999, Clubs, Ceilings and CDM, The Energy Journal,
Special Issue: The Costs of the Kyoto Protocol: A Multi-model Evaluation, pp. 177-206.
Criqui, Patrick, S. Mima and L. Viguier, 1999, Marginal abatement costs of CO2 emission reductions,
geographical flexibility and concrete ceilings: an assessment using the POLES model, Energy
Policy, 27: 585-601.
Ellerman, A. Denny and A. Decaux, 1998, Analysis of Post-Kyoto CO2 Emissions Trading Using
Marginal Abatement Curves, MIT Joint Program on the Science and Policy of Global Change,
Report No. 40, Cambridge, Mass.
Ellerman, A. Denny, H.D. Jacoby and A. Decaux, 1998, The Effects on Developing Countries of the
Kyoto Protocol and CO2 Emissions Trading, MIT Joint Program on the Science and Policy of Global
Change, Report No. 41, Cambridge, Mass.
Manne, Alan S. and R.G. Richels, 1999, On Russian ÔHot AirÕ and the Kyoto Protocol, mimeograph.
Reilly, John M., R.G. Prinn, J. Harnisch, J. Fitzmaurice, H.D. Jacoby, D. Kicklighter, P. Stone,
A.ÊSokolov and C. Wang, 1999, Multi-gas Assessment of the Kyoto Protocol, Nature, 401: 549-555.
US Department of Energy, Energy Information Administration, 1999, International Energy Outlook
1999, DOE/EIA-0484(99).
Westskog, Hege, 1996, Market Power in a System of Tradeable CO2 Quotas, The Energy Journal,
17(3):Ê85-103.
16
APPENDIX
In region j the domestic supply price pj of a quantity of abatement aj is determined by its
marginal abatement curve fj(aj). We assume that each function f passes through the origin, is
invertible, strictly increasing in abatement, and possesses a non-negative second derivative.
Every region faces an emissions reduction commitment kj and may be endowed with surplus
permits hj representing hot air.
A1. Import restrictions generate a fall in the permit price in the presence of competitive supply:
With emissions trading, jÕs net imports of emissions-reduction credits mj are defined by the
market clearance condition
mj = kj Ð hj Ð aj = kj Ð hj Ð fj Ð1(pj)
(2)
and the aggregate excess demand for permits by
åj mj = åj (kj Ð hj) Ð åj fj Ð1(pj).
(3)
Let the set of j regions in the market be partitioned into i importers and e exporters. Assuming
constancy of partiesÕ commitments and endowments of hot air, we can split equation (3) into an
aggregate demand function
å i mi = å i (ki - fi -1 ( pi ))
(4)
and an aggregate supply function
å e me = å e (ke - he - fe -1 ( pe )) .
(5)
Import restrictions imply that dmi < 0. We assume that sellers adjust their output for the
market to clear. Using the sign convention that mi < 0 and me > 0, this implies
åe dme = Ðåi dmi > 0.
With restricted trade, let pRT be the price at which the marginal cost of abatement is equalized
across exporters, such that the market clears. Linear additivity of abatement allows us to define a
function Ae of the price, so that we can rewrite equation (5) as
å e me = å e ( ke - he ) - Ae ( p RT )
implying that
åe dme = Ð Ae¢ dpRT.
The fact that the functions fe Ð1 are strictly increasing means that Ae¢ > 0, implying that dpRT < 0.
17
A2. Unrestricted (ÒfringeÓ) importers diminish lead monopsonistsÕ price-setting capability:
From equation (1), the first-order condition of the optimal restriction for a restricted importer r is
given by
dp
fr ( ar ) = p - mr
.
dar
The derivative in this expression is a total derivative that takes into account the reaction of both
exporters e and unrestricted importers u to rÕs restriction. For simplicity, consider a market in which
agents in each of these three classes of parties can be aggregated into representative individuals.
Under the assumption that the price is determined by the change in abatement of the exporter
dp
da
= fe¢ e .
dar
dar
Market clearing requires that
dae = -( dar + dau )
which may be written equivalently as
æ
dae
da ö
= -ç1 + u ÷ .
dar
dar ø
è
The solution turns on the fact that the derivative within parentheses can be stated as
dau
dp
= 1 fu¢
.
dar
dar
Combining these formulae then yields
- fe¢fu¢
dp
= - fe¢(1 + 1 fu¢) =
dar u
fu¢ + fe¢
(6)
which is negative with strictly increasing marginal abatement curves (f ¢ > 0).
This result may be compared to the equilibrium condition that obtains when there are no
unrestricted importers in the market (Øu). In this case
dae
dp
= -1 and
dar
dar
= - fe¢.
Øu
The presence of unrestricted importers therefore attenuates the response of the market
clearing price to import restrictions:
dp
dar
>
Øu
dp
.
dar u
This result embodies the intuition that attempts by unrestricted importers to expand imports to
take advantage of the fall in the permit price feed back negatively on the price reduction that can
be achieved in the market.
18
A3. With restricted trade, internal deadweight loss increases with domestic abatement, savings
from permit imports achieve a maximum, and total abatement costs achieve a minimum:
With reference to Figure 2, a restricted party r suffers a deadweight loss ABC given by
ar
DWLr =
∫f(
r
)d
− pFT (ar − a FT )
(7)
a FT
where pFT and aFT are, respectively, the permit price and level of domestic abatement under free
trade. This increment to cost is offset by the gross savings on permit imports CDEF given by
Sr = ( pFT − p(ar ))(kr − ar ) .
(8)
Differentiating these expressions yields first-order conditions
aˆr = fr −1(p FT )
and
FT
a˜r = kr − ( p − p )
1
.
p′
The interpretation of these formulae is straightforward. First, the internal deadweight loss
DWLr is minimized at the free-trade level of imports aˆr = a FT . To see this, note that the second
derivative of (7) is
d 2 DWLr
= fr′ > 0 .
2
dar
Combined with the fact the derivative of DWLr is fr – pFT > 0, this condition implies the that
deadweight loss is increasing and convex in ar. This is shown by the thin solid line in Figure 3.
Second, the gross savings on imports Sr are maximized at a level of restriction a˜r . By
equation (6) p′ < 0, implying that a˜r ∈(a FT ,kr ) . Sufficient conditions for savings to be
maximized can be found from the second derivative of (8):
d 2 Sr
= 2 p ′ − (k r − ar ) p′′
dar2
where p″ is simply the derivative of equation (6). A negative value of this expression indicates
that Sr achieves a maximum. From A2 we know that
 − fe′fu′
 f ′ + f ′ < 0 if u exists
p′ =  u e
 − fe′ < 0 if u does not exist.

Therefore, all that is required is to determine the sign of p″. When there are no unrestricted parties
19
d2 p
da
= - fe¢¢ e = fe¢¢> 0 .
2
dar
dar
When there are unrestricted parties the solution is more complicated
d2 p
-1
=
2
dar
( fe¢ + fu¢)2
é dae 2
dau 2 ù
ê da fu¢ fe¢¢+ da fe¢ fu¢¢ú .
r
ë r
û
Again, A2 facilitates substitution for
dae
da
and u to give
dar
dar
2
3
d 2 p fe¢¢fu¢ ( fe¢ + fu¢) + fu¢¢fe¢
>0
=
( fe¢ + fu¢)3
dar 2
from which the second-order condition for maximum savings is met. The result is shown by the
dashed line in Figure 3.
The difference between (7) and (8) gives the net increment to the total cost of compliance
under import restrictions:
ITCr = DWLr Ð Sr.
The first order condition for this expression gives the optimal ceiling
1
ar * = kr + ( fr - p) .
p¢
The second derivative is
d 2 ITCr
= fr¢ - 2 p¢ + ( kr - ar ) p¢¢
dar 2
which, since p¢ < 0 and p² > 0, implies that net incremental costs achieve a minimum. This is
shown by the heavy line in Figure 3.
A4. The existence and computation of equilibrium in a market for tradable permits:
At the positive free-trade permit price pFT the sum of net imports is zero, so that equation (3)
becomes
åj (kj Ð hj) Ð åj fj Ð1(pFT) = 0.
Linear additivity of regionsÕ emissions reductions allows us to express aggregate abatement as a
single function A of the price
åj fj Ð1(pFT) = A(pFT).
From the definition of fj, the functions fj-1 exist and are strictly increasing, so that A(0) = 0 and
A¢Ê> 0. If A is invertible, a closed-form solution exists for the market-clearing price as a strictly
increasing function of the commitments of all regions, net of the aggregate endowment of hot air
20
p
FT
=A
[ ∑ (k
−1
j
j
]
− hj ) .
A(·) one-one is then sufficient for a unique solution.
In general, an analytic expression for A–1 does not exist. In the paper we compute prices and
quantities in market equilibrium by finding p* such that
∑j (kj – hj) – A(p*) = 0.
This is equivalent to solving
min ∑ j ( k j − hj ) − A( p)
p>0
by iterating over prices p. A strictly increasing implies that the maximand possesses the singlecrossing property that guarantees that if a solution exists, it will be unique. It is then
straightforward to solve for the distorted equilibrium with supplementarity limits lj as a
constrained minimization:
min ∑ j ( k j − hj ) − A( p) s.t. fj-1(p) ≥ lj.
p>0
A similar framework may be used to solve for the market equilibrium in the presence of a
price-setting monopolistic exporter. For any exporting region e, the profit e it receives will be
given by revenues from permits sales net of the costs of actual abatement.
ae
e
= (ae + he ) ⋅ pe − ∫ fe ( )d
0
We assume that when a region eˆ acts as a monopolist it manipulates the price pe across all
exporters to maximize its own profit
eˆ
(pe ) = [ aeˆ (pe ) + heˆ ] ⋅ pe −
a eˆ ( p e )
∫f(
eˆ
)d
0
subject to the permit market clearing on quantity. Importers i are assumed to be acquiescent and
to respond as if competitive market conditions prevailed, equalizing their marginal costs of
abatement at a common buyer price pi that is determined by the aggregate quantity of exported
permits. The monopolistic equilibrium may then be found by solving for prices pe and pi that
simultaneously maximize eˆ ’s profits while clearing the market, which is the simple
maximization problem
max
p e , pi >0
ˆe
s.t. ∑i [ ki − fi− 1( pi )] − ∑e [ fe− 1(pe ) + he ] = 0.
21